Main

High-frequency ventilation (HFV) is advocated as a lung protective ventilation strategy for the treatment of respiratory distress syndrome (RDS) in preterm newborn infants. HFV has proven particularly useful for optimizing lung volume, reducing atelectotrauma and volutrauma, and therefore reducing injurious lung stimuli associated with bronchopulmonary dysplasia (BPD) (1–3). High-frequency jet ventilation (HFJV) and high-frequency oscillatory ventilation (HFOV) are the two main forms of HFV used in NICUs, and although there is substantial research on the optimal approach to lung volume optimization in preterm RDS using HFOV, data are limited for HFJV. To date, there is one multicenter controlled trial and one single-center controlled trial comparing the use of HFJV [with a low positive end-expiratory pressure (PEEP) strategy] to conventional mechanical ventilation (CMV) for treatment of preterm RDS (4,5). The results give conflicting information regarding the respiratory and neurological outcomes of neonates treated with HFJV. Increased adverse neurological outcomes for HFJV group in the study by Wiswell et al. (5) were attributed to hypocarbia (6). Subgroup analysis in the trial by Keszler et al. (4) suggested that the use of low PEEP was associated with an increased risk for grade III-IV intraventricular hemorrhage or periventricular leukomalacia. An important limitation of these trials is that the HFJV groups were not compared with a true “lung protective” CMV strategy. Further research assessing the pros and cons of optimized versus low PEEP in HFJV is therefore warranted.

HFJV is characterized by a (normally) fixed brief inspiratory time, passive expiration, and coupling with a conventional ventilator for provision of conventional breaths, PEEP, and bias flow. The current strategy recommended for treatment of RDS with HFJV is to commence HFJV early in the disease process with a peak inspiratory pressure (PIP) just below that being used during CMV (Bunnell Inc, HOW to use the LifePulse HFV: Seven Steps to Success, available at http://www.bunl.com/Clinical%20PDFs/New%20How%20to%20use%20the%20LifePulse%20New%2012-09.pdf; accessed October 12, 2010). The initial PEEP is set to achieve a mean airway pressure (Paw) equal to that used before commencement of HFJV. The primary method advocated for optimizing lung volume recruitment in HFJV is by incrementing PEEP until stable peripheral oxyhemoglobin saturation (SpO2) is achieved, with low-rate CMV breaths added to HFJV as a supplementary method for recruiting collapsed alveoli (algorithm provided by Bunnell Inc, Finding Optimal PEEP during HFV, available at http://www.bunl.com/Patient%20Management/OptimalPEEP.pdf; accessed October 12, 2010).

Despite the detailed guidelines provided for optimizing lung volume during initiation of HFJV, the evidence basis for this approach is limited. Although inadequate PEEP will encourage airway collapse and atelectasis and initiate the lung injury sequence, excessive PEEP promotes alveolar overdistension, impedes pulmonary perfusion, decreases venous return, and depresses cardiovascular function (7–9).

We hypothesized that a PEEP-driven lung volume recruitment protocol would enhance arterial oxygenation and ventilation without promoting lung injury during the initiation of ventilation in a preterm ovine model of RDS. Furthermore, we hypothesized that these effects would be achieved at the expense of pulmonary perfusion. We aimed to compare the effect of an adjusted PEEP strategy on pulmonary blood flow (PBF), blood gases, and lung injury with a constant low PEEP strategy in an instrumented preterm lamb model.

MATERIALS AND METHODS

All animal procedures were approved by the University of Western Australia animal ethics committee, according to the guidelines of the National Health and Medical Research Council of Australia code of practice for the care and use of animals for scientific purposes.

Animals, instrumentation, and delivery.

Single and twin-bearing date-mated merino ewes were anesthetized at 128–130 d gestation (term is ∼150 d) with intramuscular xylazine (0.5 mg·kg−1; Troy Laboratories, New South Wales, Australia) and ketamine (20 mg·kg−1; Parnell Laboratories, New South Wales, Australia) and intubated with a 7.5-mm cuffed tracheal tube (Portex Ltd., England). Maternal anesthesia was maintained with halothane in 100% O2. The fetus was exteriorized, and a right lateral thoracotomy was performed. A flow probe (4R; Transonic Systems, Ithaca, NY) was positioned around the left pulmonary artery, and a catheter was inserted into the main pulmonary artery (7). The fetus was intubated orally (4.5-mm cuffed tracheal tube; Portex Ltd.), lung fluid was suctioned, and intratracheal surfactant (100 mg·kg−1: Survanta, 25 mg of phospholipids mL−1; Abbott Laboratories, IL) was administered before cesarean section delivery of the lamb. Unventilated controls (UVCs; n = 6) were euthanized (pentobarbitone, 100 mg·kg−1 i.v.) at delivery without instrumentation.

Postnatal care.

Instrumented lambs were dried, weighed, and randomized to one of two ventilation groups: constant PEEP (PEEPconst; n = 6) and adjusted PEEP (PEEPadj; n = 6). They were commenced on HFJV according to a predetermined protocol (Fig. 1). Umbilical venous and arterial catheters were inserted. Propofol (0.1 mg/kg/min; Repose 10%, Norbrook Laboratories Ltd., Victoria, Australia) and remifentanil (0.05 μg/kg/min; Ultiva, Abbott Laboratories) were infused continuously through an umbilical vein for anesthesia and analgesia. An umbilical arterial catheter was used for continuous measurement of systemic arterial blood pressure (ABP) and intermittent sampling to assess gas exchange and acid-base balance. Rectal temperature was monitored continuously and maintained between 38 and 39°C (normothermic for newborn lambs).

Figure 1
figure 1

Constant and adjusted PEEP protocols. Lambs were ventilated at either a constant PEEP of 5 cm H2O (PEEPconst, solid line) or with stepwise increments and decrements to PEEP (PEEPadj, gray). Values are mean (SEM).

Full size image

The Oxygenation Index (OI) was calculated asformula

where FiO2 is fractional inspired oxygen concentration, Paw is mean airway pressure, and Pao2 is PO2 in arterial blood.

High-frequency jet ventilation.

HFJV (Life Pulse High-Frequency Ventilator; Bunnell Inc., Salt Lake City, Utah) coupled to a pressure-limited time-cycled infant conventional ventilator (BP 200 Infant Pressure Ventilator; Bourns Life Systems, CA) was commenced immediately following delivery. HFJV was commenced using a respiratory rate of 420 breaths per minute (7 Hz) and PIP of 40 cm H2O. HFJV PIP was adjusted to a maximum of 40 cm H2O to target moderate permissive hypercapnia (PCO2 in arterial blood (Paco2) 45–55 mm Hg). The initial FiO2 of 0.4 was adjusted to maintain SpO2 of 90–95%. Inspiratory time (tI) was fixed at 0.02 s. No conventional ventilator breaths were applied during the 2-h study period.

PEEP was maintained at 5 cm H2O in the PEEPconst group for the duration of the study. Lambs in the PEEPadj group were stabilized on a PEEP of 5 cm H2O for 10 min then PEEP was incremented at 10 min (8 cm H2O), 15 min (10 cm H2O), and 20 min (12 cm H2O). PEEP was decreased by 2 cm H2O at 35 min and 60 min and then maintained at 8 cm H2O from 60 min until euthanasia at 120 min.

Continuous measurements of PBF, pulmonary artery blood pressure (PAP), and systemic ABP were processed via calibrated pressure transducers (Maxxim Medical, TX). Data were amplified and stored on a digital data acquisition system (Powerlab 8SP; ADInstruments, New South Wales). Pulmonary wave form analysis was performed at regular time points as described previously (10). Pulsatility Index, a measure of downstream resistance to blood flow, was calculated as (peak systolic flow − minimum flow after diastolic flow)/mean peak systolic flow over five consecutive cardiac cycles.

Tidal volume (VT) was measured continuously using an electronic flowmeter (Florian, Acutronics, CH). Ventilator settings (respiratory rate, PIP, PEEP, Paw, delta P (ΔP), servo pressure, and ti) were recorded at intervals. After final measurements were obtained, the FiO2 was increased to 1.0 for 2 min, the tracheal tube was occluded for 3 min, and the lamb was euthanized (100 mg·kg−1 pentobarbitone i.v.).

Postmortem.

The lung was exposed by thoracotomy, and an in situ deflation pressure volume curve was obtained (11). The right upper lung lobe was inflation fixed (30 cm H2O) in formalin and samples of the right lower lobe were snap frozen for molecular analyses. Bronchoalveolar lavage (BAL) was performed on the left lung for cytology and protein analysis by the Lowry method. Differential cell counts were performed on cytospin samples of the BAL fluid stained with Diff-Quik (Fronine Lab Supplies, New South Wales, Australia).

Total RNA was isolated from 30 mg of homogenized lung tissue using the RNeasy Mini kit (Qiagen, USA) according to the manufacturer's instructions. The contaminating genomic DNA was removed by an on-column DNaseI digestion performed using the DNaseI digestion kit (Qiagen). One microgram of RNA was then reverse transcribed into cDNA in a 20-μL reaction with QuantiTect Reverse Transcription Kit (Qiagen). The primers used for amplifying IL 1β, IL-6 (12) and IL-8, early growth response (EGR) 1, connective tissue growth factor (CTGF), and cysteine rich 61 (CYR 61) were described previously (13). Amplification and detection of specific products were conducted on the Rotor-gene 3000 real-time PCR system (Corbett Life Science) with the published cycle profiles using Rotor-gene SYBR Green PCR Kit (Qiagen) following the manufacturer's instructions. The expression levels of genes of interest were normalized into 18S RNA (14) using the 2−ΔΔCT method (15) and presented as expression ratio relative to the UVC group.

Myeloperoxidase (MPO) activity was measured spectrophotometrically using methods described by Faith et al. in 2008 (16) and McCabe et al. in 2001 (17). The MPO activity was normalized to the total protein content of the tested samples. Activity was expressed as units of MPO activity per milligram of protein, where one unit of MPO was defined as the amount needed to degrade 1 μmol of hydrogen peroxide per minute at room temperature.

Statistical analyses.

For comparison of two groups of ventilated animals at specific time points, a Mann-Whitney rank-sum test (nonparametric data) or a t test (parametric data) was used. Comparisons of two ventilated groups against the UVCs used one-way ANOVA. A two-way repeated measure ANOVA was used to determine the effect of PEEP on PBF and Pulsatility Index and the effect of time on PIP, ΔP, Paw, and servo pressure. Data in the text and legends are expressed as mean (SEM) or median (25th percentile, 75th percentile) unless otherwise stated. Analyses were performed using SigmaStat (version 3.5; Systat Software Inc, Chicago, IL), with p < 0.05 considered statistically significant.

RESULTS

Baseline characteristics of lambs in each group were not different (Table 1).

Table 1 Baseline characteristics
Full size table

Ventilator settings and lung mechanics.

Changes in Paw reflected the different PEEP protocols (Fig. 2A) but as PIP decreased in both groups, ΔP also decreased (Fig. 2B). The VT was higher in PEEPadj group (pooled time points: p < 0.01) despite a significantly lower ΔP (Fig. 2C). Servo pressure was lower (p = 0.026) in the PEEPadj group at 120 min (Table 2). Pressure-volume curves showed a higher volume achieved per unit of pressure for the PEEPadj lambs compared with the PEEPconst lambs (p = 0.003 at 40 cm H2O) and for both of the ventilated groups compared with the UVCs (Fig. 2D).

Figure 2
figure 2

Ventilator settings and lung mechanics. (A) Mean airway pressure (Paw), (B) delta P (ΔP), (C) tidal volume, and (D) deflation limb of the postmortem in situ pressure-volume curves. ⋄, PEEPconst; •, PEEPadj; , unventilated control.*p < 0.05 compared with PEEPconst and **p < 0.05 compared with unventilated control. Values are mean (SEM).

Full size image
Table 2 Cardiovascular and respiratory physiological measurements
Full size table

Oxygenation.

The target SpO2 of 90–95% was achieved in both groups within the first 10 min and was maintained throughout the ventilation period. FiO2 and OI were lower in the PEEPadj group compared with the PEEPconst lambs from 45 min until the end of the ventilation period (Fig. 3A and B).

Figure 3
figure 3

Oxygenation. (A) FiO2, (B) OI over the 120-min ventilation period. ⋄, PEEPconst; •, PEEPadj. *p < 0.05 compared with PEEPconst. Values are mean (SEM).

Full size image

Blood gases.

The target Paco2 was achieved within 10 min in both groups and was maintained between 45 and 55 mm Hg throughout the ventilation procedure (Table 2). The pH, arterial lactate concentration, and base excess (BE) were comparable between the two groups for the duration of the ventilation period (Table 2).

Hemodynamic consequences of different PEEP protocols.

PBF increased over the first 10 min of ventilation in both groups, and there were no significant differences between the groups at any time point between 10 and 120 min. PBF decreased by ∼8% in the PEEPconst group and by ∼48% in the PEEPadj group between 10 and 120 min (p = 0.507 and p = 0.026, respectively; Fig. 4A). After an initial decrease during the transition from fetal to neonatal circulation, the Pulsatility Index increased in both groups over time and was significantly higher in the PEEPadj group over much of the first 60 min (Fig. 4B). The pulmonary and systemic ABPs were not different between the groups at any time point (Table 2). End-diastolic and end-systolic PBFs were significantly decreased at 15 min in the PEEPadj group (p < 0.001 for each parameter). For all pulmonary artery variables, the difference between PEEPconst and PEEPadj was temporary (Fig. 4C and D).

Figure 4
figure 4

Pulmonary perfusion. (A) PBF, (B) Pulsatility Index over the 120-min ventilation period, (C) End-diastolic blood flow, (D) End-systolic blood flow. ⋄, PEEPconst; •, PEEPadj. *p < 0.05 compared with PEEPadj and **p < 0.05 time 120 min compared with time 10 min. Values are mean (SEM).

Full size image

Lung injury.

BAL fluid protein concentration was higher in ventilated groups than in the UVCs, but no difference between PEEPadj and PEEPconst was observed. The cell populations of BAL fluid did not differ between the groups (Table 3).

Table 3 Postmortem inflammatory markers
Full size table

In lung tissue, the expression of IL-1β, IL-6, IL-8, CTGF, and EGR1 mRNA was increased in the PEEPconst group compared with UVCs, whereas CYR 61 expression was significantly increased in the PEEPadj group compared with UVCs. The expression of IL-1β, IL-6, EGR1, and CTGF was greater in the PEEPconst group compared with PEEPadj. There was no difference in MPO activity (Table 3) between each of the three groups.

DISCUSSION

Alveolar recruitment during HFJV can be achieved by delivering CMV breaths or adjusting PEEP, or both. To examine the role of PEEP for alveolar recruitment during HFJV, this study aimed to compare the effect of an adjusted versus a constant PEEP protocol on oxygenation, ventilation, pulmonary hemodynamics, and lung injury during HFJV. We showed that lambs in the PEEPadj group had better oxygenation and lung compliance but decreased PBF compared with PEEPconst lambs. Markers of lung injury were higher in the PEEPconst group in this short-term study.

The correlation between Paw and oxygenation during HFJV is well understood. As expected by study design, the Paw of the PEEPadj group was significantly higher than the PEEPconst group throughout the ventilation period. The benefit of higher Paw was evidenced by lower FiO2 requirements from as early as 45 min, supporting the concept that a high Paw strategy enhances arterial oxygenation, decreasing the FiO2 required during ventilation (18). In a clinical setting, PEEP is increased and FiO2 is decreased until the SpO2 or Pao2 plateaus or falls. When FiO2 is stable at 0.21, optimal PEEP has been achieved (19). The OI describes the relationship between FiO2, Paw, and Pao2: a lower value implies better arterial oxygenation (20). After the initial decrease in OI within the PEEPconst group, the OI increased over the remainder of the study. In the PEEPadj group, it remained relatively low and stable despite the higher Paw, suggesting more efficient gas exchange due to either airway stenting or increased gas-exchange surface due to effective volume recruitment.

Servo pressure is the automatically controlled driving pressure of the ventilator and changes in response to altered monitored airway pressure to ensure that the ventilator will continue to deliver inspiratory gas to meet the set PIP (21). Servo pressure increases as PIP increases, with increased lung and/or chest wall compliance, a decrease in airway resistance, or if there is an air-leak or circuit disconnection. A decrease in servo pressure however, indicates either a reduction in set PIP, or worsening compliance and increased resistance to gas flow, obstruction of the tracheal tube, tension pneumothorax, and the requirement for suctioning or right mainstem bronchus intubation (www.bunl.com/ServoSlidesNEW.html). Our finding of a lower servo pressure throughout most of the study period in PEEPadj lambs, despite evidence of increased lung compliance, likely reflects the reductions in PIP required to maintain moderate permissive hypercapnia and to avoid overventilation.

The hemodynamic consequences of the ventilation protocol were assessed by measurement of PBF, PAP, systemic ABP, pulse rate, and calculation of Pulsatility Index. Pulsatility Index is directly related to resistance to blood flow. An increase in intrathoracic pressure during any kind of positive pressure ventilation impacts on venous return, cardiac output, right ventricular end-diastolic volume, PBF, and pulmonary vascular resistance (22). Increased PEEP reduces PBF during conventional ventilation of very premature lambs by increasing pulmonary vascular resistance (PVR) (8). A similar decrease in PBF in response to Paw-driven alveolar recruitment maneuvers during HFOV of up to 69.3% is also reported (7). Importantly, during HFOV, the fall in PBF persisted after Paw was decreased following recruitment (7). The mechanism causing a fall in PBF as PEEP (and consequently also Paw) is increased may include an increase in the alveolar capillary transmural pressure causing capillary compression (8,23). The noncompliant immature lung may be less susceptible to capillary compression than a mature lung (8,23). Factors that increase lung compliance (e.g. antenatal corticosteroids, exogenous surfactant, and lung volume recruitment) may increase the sensitivity of PBF to changes in airway pressure (23). We observed previously that incrementing PEEP to 12 cm H2O during conventional ventilation improved oxygenation at the expense of PBF (8). Improved oxygenation at the expense of PBF was also observed previously using an incremental Paw strategy during HFOV (7).

The fall in PBF in the PEEPadj group reversed rapidly as PEEP was initially decreased, suggesting the impact of a PEEP recruitment strategy on PBF during HFJV was not sustained. Furthermore, the impact of increased PEEP on end-diastolic and end-systolic blood flows coincided with the initial increase in PEEP from 5 to 8 cm H2O at 10 min. The difference between the two groups was temporary, and despite further increases in PEEP in the PEEPadj group, the blood flow variables did not remain significantly different. The maintenance of a higher Paw in the PEEPadj group likely contributed to the continued slow decline in PBF in this group and returning to PEEP of 5 cm H2O (instead of 8 cm H2O) may have been prudent. Although the inequality of Paw strategy does not allow direct comparison of PBF between the PEEPadj group, and previous observations of a delayed recovery of PBF after a stepwise Paw recruitment strategy during HFOV (7), differences between effect of HFJV and HFOV are plausible theoretically. Lung volume recruitment is most often achieved during HFOV by incrementing and decrementing Paw. Although incrementing and decrementing PEEP during HFJV may also alter Paw, the subject spends substantially longer at PEEP than at the PIP, with potential for less negative impact on PBF. A direct comparison of HFJV and HFOV in a controlled setting using clinically defined and comparable recruitment strategies is warranted to determine the extent and duration of their impact on pulmonary hemodynamics.

Lung inflammation is a prelude to lung injury. We examined inflammatory markers that we anticipated would be increased at 120 min in response to lung injury (13,24–26). There was a clear increase in lung injury across the range of inflammatory markers for the PEEPconst group, compared with the UVCs and for the PEEPconst group compared with PEEPadj. These findings support our hypothesis regarding reduced lung injury with PEEP recruitment of the lung during HFJV.

MPO activity has been shown to correlate with IL6 expression (16), but our results did not show a clear relationship between these variables. The MPO activity was not different between the groups, despite a trend for an increase in the ventilated groups. It is possible that this is a result of maternal anesthesia, and this finding in itself warrants further investigation. It is also possible that a Type II statistical error as a result of the small group sizes prohibited a demonstrable difference between the groups.

There are a number of limitations to our study. We wanted to examine the physiological changes associated with each PEEP alteration and required at least 10 min to accommodate and document any changes. Consequently, the adjusted PEEP protocol that we studied does not reflect standard clinical practice given that the 60-min period for increasing and decreasing PEEP is considerably longer than optimally used in a clinical setting. This period of time at a higher PEEP may have attenuated our lung injury results and potentially masked significant differences between the ventilated groups. Second, surfactant was administered to the lambs before the commencement of ventilation. This practice may not be achieved routinely in a clinical setting. However, our goal was to isolate the effect of PEEP and standardize and optimize all other aspects of care. Third, the lambs in our study were anesthetized and underwent an invasive surgical procedure. The hemodynamic effects of anesthesia combined with the physical impact of instrumentation on lung inflation are likely to impact physiological outcomes. Finally, the flowmeter used for measuring VT slightly overestimates tidal volume at 7 Hz (27). However, as HFJV frequency was constant throughout the study, we would not expect this to affect comparisons between the ventilatory groups.

In conclusion, adjusted PEEP during HFJV improves oxygenation and lung compliance and reduces ventilator requirements despite reducing pulmonary perfusion. The majority of markers of injury were higher when PEEP was constant during HFJV. Evaluation of PEEP recruitment maneuvers in human patients is indicated to explore the efficacy of the technique in the target patient population.